This invention relates to a process for optimizing ligand-target molecule system selection and using a selected system for preparing large scale quantities of macromolecules in particularly pure form.
The production of an array of ligands to select a ligand or a set of ligands for use in a macromolecule purification scheme has traditionally been by organic synthesis methods. Such methods include solid-phase and liquid-phase synthesis. Solid-phase methods are capable of producing longer ligands than liquid-phase methods thus making solid-phase methods preferable. However, ligands produced using either of these methods are difficult to identify and amplify.
Ligand synthesis in a biological system is advantageous over organic synthesis because of the ability to identify and amplify large quantities of ligands. Until recently, biological system synthesis of an array of ligands, such as an epitope library, was limited. However, random peptide libraries that take advantage of the ability of filimentous phage coat protein gene pIII to accept and express foreign DNA on its surface have been described as being useful to identify millions of potential ligands quickly. See, e.g., Scott, J. K. and Smith, G. P., Science, 449, 386-390 (1990), Devlin, J. J., et al., Science, 449, 404-406 (1990), Cwirla, S. E. et al., Proc. Nat. Acad. Sci. USA, 87, 6378-6382 (1990) and U.S. Pat. No. 5,223,409.
Historically, ligand-target molecule system selection for macromolecules have relied upon secondary indication methods to determine appropriate ligands. Such methods include RIA, ELISA, and biotin-avidin complex formation assays. Although these methods identify ligand-target molecule systems that have an unusually high association constants, it is only with subsequent rounds of screening that one skilled in the art can identify ligand-target molecule systems with sufficient binding characteristics for use in subsequent macromolecule purification. These techniques have other deficiencies including the ability to produce false positives, being time consuming and lacking the ability to differentiate between active and nonactive macromolecules during purification.
Surface plasmon resonance (SPR) has been known for quite some time, Kreetschmann, E., and Raether, H., Z. Naturforsch. A23, 2135 (1968). However, it was not until recently that the use of SPR to study ligand-target molecule interactions was described, Karlsonn, R., et al., J. Immunol. Methods, 145, 249 (1991).
It is well known that affinity chromatography is an effective purification approach that exploits a macromolecule""s biological function. Most macromolecules possess active sites that perform unique functions. These active sites are involved in the recognition and the catalysis of selected small molecules or restricted regions of other macromolecules. It is the property of recognition upon which the principles of affinity chromatography have been developed. The fundamental requirement of affinity chromatography is that the comparative rate constants reflect reasonable affinity, and that the qualitative nature of the ligand and target molecule reflect reasonable stereochemical specificity.
It is desirable that an interacting ligand-target molecule system be chosen such that the ligand-target molecule complex is not chemically altered as a result of the interaction. Many nonenzymatic interacting systems do not exhibit such chemical alteration and are, therefore, ideally suited for affinity chromatography purification. Such interacting systems theoretically include antigens, antibodies, vitamin and drug binding proteins, biological receptors, and transport proteins.
It is also desirable that the interacting ligand-target molecule system be chosen such that the target molecule binds sufficiently fast to the ligand and that the ligand-target molecule system exhibit a sufficiently slow dissociation, thereby allowing large quantities of the target molecule to couple with the ligand without significant loss of target molecules before elution. Following these parameters it is possible to increase the purity and amount of target molecule ultimately recovered.
The aforementioned techniques are themselves individually known. However, the combination of these techniques to identify ligand-target molecule systems with specific association and dissociation constants for subsequent purification of target molecules and the subsequent purification of target molecules using identified ligand-target molecule systems is not known.
The following terms are used herein according to the definitions.
The terms xe2x80x9cprotein,xe2x80x9d xe2x80x9cpeptide,xe2x80x9d xe2x80x9coligopeptide,xe2x80x9d and xe2x80x9cpolypeptidexe2x80x9d and their plurals will be used interchangeably to refer to chemical compounds having amino acid sequences of five or more amino acids. xe2x80x9cAmino acidxe2x80x9d refers to any of the 20 common amino acids for which codons are naturally available and are listed in the table of amino acids.
As used herein, all amino acid three letter and single letter designations conform to those designations which are standard in the art, and are listed as follows:
When any variable, e.g., SPNE, occurs more than one time in any constituent, its definition on each occurrence is independent of its definition at every other occurrence. Also, combinations of substituents, variables, or both are permissible only if such combinations result in stable compounds.
The present invention overcomes the problems of previous methods for identifying specific ligands for the purification of macromolecules. One embodiment of the present invention is to prepare RFPs selected as SPNEs of phage libraries that bind to a specific target molecule. Another embodiment of the present invention is to determine, in real time, the association and dissociation constants for the RFPs and the target molecule of choice and then select RFPs that meet predetermined binding characteristics in order to optimally purify the target molecule. A further embodiment of the present invention is to provide sufficient qualities of identified RFPs for target molecule purification. Yet a further embodiment of the present invention is to use at least one of the identified RFPs as a ligand to obtain large quantities of a target molecule in an active and particularly pure form in an affinity chromatography purification scheme. These and other objects will become apparent to those skilled in the art in the following disclosure.
It is now a relatively straight forward technology to prepare cells expressing a foreign gene. Such cells act as hosts and include bacteria, yeast, fungi, plant cells or animal cells. Expression, vectors for many of these host cells have been characterized and are used as starting materials in the construction, through conventional recombinant DNA techniques, of vectors having a foreign DNA insert of interest. Any DNA is foreign if it does not naturally derive from the host cells used to express the DNA insert. The foreign DNA insert may be expressed on extra-chromosomal plasmids after integration in whole or in part of the host cell chromosome(s), or may actually exist in the host cell as a combination of more than one molecular form. The choice of host cell and expression vector for the expression of a desired foreign DNA largely depends on availability of the host cell and how fastidious it is, whether the host cell will support the replication of the expression vector and other factors readily appreciated by those of ordinary skill in the art.
The technology for recombinant prokaryotic expression systems is well developed and reproducible. A typical host cell is E. coli. The technology is illustrated by treatises such as Wu, R (ed), Meth. Enzymol., 68 (1979) and Maniatis, T. et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor (1982) and updates thereof.
The foreign DNA insert of interest is any DNA sequence coding for a SPNE (or fragment thereof of at least 5 amino acids in length) of the present invention, including any synthetic sequence with this coding capacity or any such cloned sequence or combination thereof. For example, SPNE peptides coded and expressed by an entirely recombinant DNA sequence is encompassed by this invention.
Vectors useful for constructing prokaryotic expression systems for the production of recombinant SPNE include the DNA sequence for SPNE, fragment or variant thereof, operatively linked thereto with appropriate transcription activation DNA sequences, such as a promoter, an operator or both. Other typical features include appropriate ribosome binding sites, termination codons, enhancers, terminators and replicon elements. These additional features can be inserted into the vector at the appropriate site or sites by conventional splicing techniques such as restriction endonuclease digestion and ligation.
Yeast expression systems, which are one variety of recombinant eukaryotic expression systems, generally employ Saccharomyces cerevisiae as the species of choice for expressing recombinant proteins. S. cerevisiae and similar yeasts possess well known promoters useful in the construction of yeast expression systems, including but not limited to GAP491, GAL10, ADH2 and alpha mating factor.
Yeast vectors useful for constructing recombinant yeast expression systems expressing SPNE include, but are not limited to, shuttle vectors, cosmids, chimeric plasmids, and those having sequences derived from 2-micron circle plasmids.
Insertion of the appropriate DNA sequence coding for SPNE, fragment or variant thereof, into these vectors will, in principle, result in a useful recombinant yeast expression system for SPNE where the modified vector is inserted into the appropriate host cell, by transformation or other means.
Recombinant mammalian expression systems are another means of producing the recombinant SPNE of this invention. In general, a host mammalian cell can be any cell that has been efficiently cloned in cell culture. Host mammalian cells useful for the purpose of constructing a recombinant mammalian expression system include, but are not limited to, Veer cells, NIH3t3, COS, murine C127, NSO or mouse L cells. Mammalian expression vectors can be based on virus vectors, plasmid vectors which may have SV40, BPV or other viral replicons, or vectors without a replicon for animal cells. Detailed discussions on mammalian expression vectors can be found in the treatises of Glover, D. M. (ed.) xe2x80x9cDNA Cloning: A Practical Approach,xe2x80x9d IRL (1985), Vols. I and II.
Phage epitope libraries are unusually versatile vehicles for identifying new antigens or ligands. The ability to obtain a phage epitope library that bind to antibodies and other receptors has been described by the following: Scott, J. K. and G. P Smith. xe2x80x9cSearching for Peptide Ligands with an Epitope Libraryxe2x80x9d. Science, 249:386-390, (1990); Devlin, J. J., L. C. Panganiban, and P. E. Devlin. Random Peptide xe2x80x9cLibraries: A Source of Specific Protein Binding Moleculesxe2x80x9d. Science 249:404-406, (1990); Cwirla, S. E., et al., xe2x80x9cPeptides on phage: A vast library of peptides for identifying ligandsxe2x80x9d. Proc. Natl. Acad. Sci. USA, 87:6378-6382(1990); and U.S. Pat. No. 5,223,409. Typically, the phage has inserted into its genome a small, randomly generated DNA sequence, e.g., 45 base pairs, which will generate exposed oligopeptide surfaces in the mature phage. Mixing a library of such mature phage with a screening antibody of desired specificity, followed by separation of bound from unbound phage, allows the opportunity to clone and sequence the bound phage. A conventional example of a phage epitope library is the filamentous phage fd and its gene III coding for minor coat protein pIII.
A rapid method of constructing a phage library containing random fifteen amino acid epitopes has been described by Scott, J. K. et al., Science 249, 386 (1990). This protocol utilizes synthetic 110 BP BglI fragments which were prepared containing degenerate coding sequence (NNK)15, wherein N stands for an equal mixture of G, A, T and C, and K stands for an equal mixture of G and T. The library is constructed by ligating the synthetic 110 bp BglI fragments in phage. fUSE5 and transfecting E. coli cells with the ligation product by electroporation. The resulting phage oligopeptide epitope library has a complexity of approximately 40xc3x97106 different epitopes.
Standard and conventional methods exist for rapid and accurate synthesis of long peptides on solid-phase supports. Solution-phase synthesis is usually feasible only for smaller peptides.
Synthesis on solid-phase supports, or solid-phase synthesis, is most conveniently performed on an automated peptide synthesizer according to, e.g., Kent S. et al., xe2x80x9cModem Methods for the Chemical Synthesis of Biologically Active Peptides,xe2x80x9d in Alitalo, K. et al., (eds.). Synthetic Peptides in Biology and Medicine, Elsevier 1985, pp. 29-57. Manual solid-phase synthesis may be employed instead, by following the classical Merrifield techniques, as described, e.g., in Merrifield, R. B., J. Am. Chem. Soc. 85, 2149 (1963), or known improvements thereof. Solid-phase peptide synthesis may also be performed by the Fmoc methods, which employs very dilute base to remove the Fmoc protecting group. Segment synthesis-condensation is a further variant of organic synthesis of peptides as within the scope of the techniques of the present invention.
In organic synthesis of peptides, protected amino acids are condensed to form amide or peptide bonds with the N-terminus of a growing peptide. Condensation is usually performed with the carbodiimide methods by reagents such as dicyclohexylcorbodiimide or N-ethyl, N1 (xcex-methylaminopropyl) carbodiimide. Other methods of forming the amide or peptide bond include, but are not limited to, synthetic routes via and acid chloride, azide, mixed anhydride or activated ester. Common solid-phase supports include polystyrene or polyamide resins.
The selection of protecting groups of amino acid side chains is, in part, by the amino acid and the peptide components involved in the reaction. Such amino-protecting groups ordinarily employed include those which are well known in the art, for example, urethane protecting substituents such as benzyloxycarbonyl (carbobenzoxy), p-methoxycarbobenzoxy, p-nitrocarbobenzoxy, t-butyloxy-carbonyl, and the like. It is preferred to utilize t-butoxy-carbonyl (BOC) for protecting the xcex5-amino group, in part because the BOC protecting group is readily removed by relatively mild acids such as trifluoroacetic acid (TFA) or hydrogen chloride in ethyl acetate.
The OH group of Thr and Ser may be protected by the Bzl (benzyl) group and the xcex5-amino group of Lys may be protected by the isopropoxycarbonyl (IPOC) group or the 2-chlorobenxyloxycarbonyl (2-C1-CBZ) group. Treatment with hydrogen fluoride or catalytic hydrogenation are typically employed for removal of IPOC or 2-C1-CBZ.
For preparing cocktails of closely related peptides, see, e.g., Houghton, R. A., Proc. Nat. Acad. Sci. USA, 82, 5131 (1985).
It is highly desirable to select desired phage epitopes for subsequent screening from a phage library containing about 40xc3x97106 epitopes. Several methods are available for screening such a large library; including, but not limited to, biotin-avidin complex formation. The applicants have employed a method for screening phage epitope libraries that involves selection of epitopes by binding a phage expressing a foreign protein on its coat to a solid-phase supported antibody. This method is useful for virtually any antibody, i.e., polyclonal or monoclonal or collection of monoclonals thereto. Any antigen can be screened. The screening method employed by the present invention is illustrated by HIV antigens screened with an HIV specific broadly neutralizing antibody (hereinafter 447 antibody).
Methods of producing 447 antibody may be found in WO 93/08216 (Apr. 29, 1993). More specifically, 447 antibody is a monoclonal antibody identified from a human patient. Human blood specimens donated from HIV-1 positive individuals were the source of peripheral B cells expressing neutralizing antibodies. These cells were immortalized by Epstein-Barr virus (EBV) infection, then individual B cell clones were screened for their ability to secrete antibody which bound a peptide sequence representing the V3 loop of HIV-1 strain MN in a solid phase ELISA format. B cell clones positive in this assay were subsequently stabilized by their fusion to the SHM-D33 cell line (a murinexc3x97human heterohybridoma, ATCC CRL 1668). Resultant B cell-heterohybridoma clones were screened for their production of antibody which recognizes the MN V3 loop peptide in a solid-phase ELISA. These procedures establish the criteria for identification and isolation of stable human antibody-producing cells wherein the antibody produced is potentially useful for development into a substance for treatment prophylatically in cases of suspected HIV-1 exposure, and therapeutically in HIV-1 positive individuals.
Screening of mature phage expressing a foreign gene involves two separate methods. First, selection of desired phage epitopes with a solid-phase supported antibody of any desired specificity. The second method, which is optional, relates to identification of desired phage epitopes by antibody lifts.
A. Selection
Selection of desired phage epitopes in a phage epitope library is performed as follows. An essentially pure preparation of antibody is adsorbed or otherwise attached to a solid-phase support, hereinafter also referred to as solid-phase supported Ab. The most preferred embodiment is monoclonal antibody adsorbed to polystyrene beads large enough to be picked up with tweezers, e.g., with a diameter of about 0.25 inch. Such large beads contribute to the ease of subsequent washing steps. Other embodiments include any solid-phase adsorbent for antibody, or any plastic, or glass bead or polysaccharide gel, e.g., SEPHAROSE. Polysaccharide gels are typically covalently conjugated to purified antibody by, e.g., cyanogen bromide.
Incubation of the solid-phase supported Ab with BSA, milk solids or other reagent for blocking non-specific interactions is preferable before selection. The presence of low levels of mild or nonionic detergent is desirable, e.g., 0.5% (v/v) of one or more in the polyoxyethylene (20) sorbitan monoleate series (TWEEN), octylglucopyranoside or Nonidet NP-40. It is apparent to those skilled in the art how to adjust the conditions for coating with such blocking agents.
An appropriate density of antibody should be determined by titration. Applicants have successfully performed selection with a density of about 0.1 xcexcg 447 antibody/cm2 on polystyrene beads (d=0.25 inch). Densities in the lower range select high affinity epitopes because of the reduced incidence of multivalent binding by the antibody to the multiple copies of the epitope on the phage tip. It is apparent to those skilled in the art how to determine the most suitable density for an antibody preparation, by monitoring the bound phage population. As a general rule, a manageable complexity of bound and eluted phage ranges from about 5xc3x97103 to about 105 phage.
Throughout the selection method described below, a wide variation in incubation times, washing times, temperature and pH is covered. It is apparent to those skilled in the art that, given a particular incubation or washing step, a suitable set of variant reaction conditions can be readily ascertained. Applicants have found that temperature and pH are critical in the stringent selection of high affinity epitopes, e.g., temperatures exceeding about 70xc2x0 C. at neutral pH or exceeding about 38xc2x0 C. at about pH 4.0, are lethal to the phage. Aside from the critical parameters of temperature and pH, the typical buffer may contain a non-specific blocking agent such as bovine serum albumin (BSA) or milk solids, as well as low levels of a nonionic detergent. For example, TTBS (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% (v/v) TWEEN-20) in about 1 mg/ml BSA is typical.
Solid-phase supported antibody is first incubated with the epitope phage library to effect binding of the phage epitopes to the antibody. It is preferred to use enough phage to vastly exceed the library complexity, e.g., 1011 phage, which is 1000 fold more than its complexity of 108. Incubation between about 4xc2x0 C. and about 65xc2x0 C., for at least 10 minutes is performed. Applicants typically incubate overnight at about 4xc2x0 C. Alternatively, a one hour incubation at about 37xc2x0 C. will select epitopes binding at a fast association rate. Incubation conditions are subject to a wide range of variations, as discussed above, but a neutral buffer containing a non-specific blocking agent is preferred, e.g., TTBS, 1 mg/ml BSA.
Washing the phage-bound solid-phase supported antibody to remove unbound phage is carried out in a variety of conditions, depending on the desired stringency. Generally, the higher the desired stringency, the higher the temperature conditions of washing, up to about 70xc2x0 C. in some conditions.
For high stringency selection, washing of the bead with bound phage is carried out by washing from about 3 to about 20 times in buffer, e.g., T.T.B.S., at neutral pH at about 65xc2x0 C. without blocking agent (hereinafter 65xc2x0 C. wash). Low-affinity phage epitopes are then eluted by washing one or more times by brief (about 2 to 5 minutes) immersion in a mildly acidic buffer without blocking agent (at about pH 3.0 to about pH 5.0 with pH 4.0 being preferred) at about ambient temperature or between about 4xc2x0 C. and about 37xc2x0 C. (the pH 4.0 wash). The pH 4.0 wash is optional in high stringency selection, but it cannot be completely combined with the 65xc2x0 C. wash. For example, phage die in about pH 4.0 buffer at about 65xc2x0 C.
High stringency selection may be enhanced by lowering the antibody density on the bead or other solid-phase support. In this case, lowering the probability that a given phage will bind more than one antibody molecule selects for higher affinity epitopes.
Lower stringency selection is performed instead by washing about 3 to about 20 times at neutral pH at about room temperature. A pH 4.0 wash may optionally follow.
Elution of high affinity epitopes is the next step (hereinafter the pH 2.0 elution) for both high and low stringency selection. Phage bound to solid-phase supported antibody are incubated briefly (about 1 to about 15 minutes) in a low pH buffer in about 0.1-10 mg/ml BSA or other non-specific binder. The buffer pH can vary from about 2.3 to about 1.0, with 2.2 being preferred. Temperature conditions range from about 4xc2x0 C. to about 37xc2x0 C., with about room temperature being desirable. Preferred buffered conditions are 0.1 N glycinexe2x80xa2HCL, pH 2.2, 1 mg/ml BSA at about room temperature.
After the pH 2.0 elution, the eluted solution containing the phage is neutralized by standard and well-known techniques. The eluted phage are grown in infectable E. coli, e.g., tet+ phage are grown in tetxe2x88x92 E. coli, e.g., K91K cells, on media containing tetracycline.
This concludes one cycle of selection, either at high stringency or low stringency. Repetition of the cycle is often found advantageous, as it lowers the complexity of eluted phage to manageable quantities (less than about 105). Repeating the cycle about 2-10 times, preferably 3-5 times, is found most practical. It will be apparent to those skilled in the art that indicated variations are readily performed and evaluated, such as switching from high stringency to low stringency on one or more cycles of selection, or changing the buffer or its pH.
B. Identification With Antibody Lifts
A second selection method which may be used in addition to the method described above is to identify, using antibody lifts, those clones with desired epitopes. The principle is to place an overlay in culture plates of cells infected with selected phage epitopes, remove the overlay, block the overlay, incubate the blocked overlay with desired antibody, label the bound antibody, and locate on the original culture plate those colonies that bind the antibody. Versions of this overlay technique that differ from the present method exist in the literature. Methods known are typically adopted for use with plaque formers, unlike the present invention. See, e.g., Young, R. A. et al., Proc Natl. Acad Sci., 80, 1194 (1983); Ausubel, F. M. et al., (eds.), xe2x80x9cScreening Recombinant DNA Libraries,xe2x80x9d in Current Protocols in Molecular Biology, Chapter 6, Greene 1989; and Davis, L. G. et al., Basic Methods in Molecular Biology, pp. 214-215, Elsevier 1986.
Plates having epitope phage-infected colonies are incubated until the colonies are sufficiently large, i.e., between about 1 mm and about 4 mm in diameter, yielding mature plates.
Mature plates are overlaid with a disk that binds proteins. The disc is typically nitrocellulose, but it may also be IMMOBILON P, cellulose acetate and the like. The disk is immediately removed and subjected to further treatment.
Blocking the overlay or disk is first performed to eliminate or substantially reduce the background of non-specific interactions. Useful blocking agents include BSA, milk solids and similar proteinaceous preparations. One preferred embodiment for this blocking step is soaking each disk for 4 hours in TABS, 10% evaporated milk, at room temperature. A preferred range is incubation for at least 2 hours, in a buffer near neutrality (about pH 5.0-8.0) containing about 0.1% (v/v) to about 1.0% (v/v) neutral detergent, in about 1% to about 20% blocking agent, within a temperature range of about 4xc2x0 C. to about 80xc2x0 C.
Washing the blocked disks to remove excess blocking agent follows, and is carried out in a buffer lacking the blocking agent. One preferred embodiment for this washing step is soaking each disk two or three times in TTBS, pH 7.3-7.5, at room temperature. A preferred range of conditions is soaking for at least 10 minutes, in a buffer with a pH that does not destroy antibody (5.0-8.0), containing 0.1% (v/v) to 1.0% (v/v) neutral detergent, within a temperature range of about 4xc2x0 C. to about 80xc2x0 C.
Contacting the disk with screening antibody follows. One preferred embodiment is incubating the washed disks overnight at 4xc2x0 C. with gentle rocking, in TTBS, 1% evaporated milk, 0.5 to 1.0 xcexcg/ml antibody. A preferred range of conditions is incubating the disks for at least 4 hours, within a temperature range of between about 4xc2x0 C. and about 65xc2x0 C., in buffer near neutrality containing about 0.1 % (v/v) to about 1.0% (v/v) neutral detergent, in 0.1% to 5% blocking agent, and 0.1 to 5 xcexcg/xcexcl antibody.
A second series of washes are performed, here to remove excess or unbound antibody. One preferred embodiment is soaking each disk four times in TTBS for 20 minutes at room temperature with gentle rocking. Preferred ranges of conditions are at least 2 soaks in buffer without blocking agents at a pH near neutrality (6.0-8.0), for 5 minutes to 1 hour, between about 10xc2x0 C. and 45xc2x0 C.
The resulting washed disks having bound antibody are treated with a labeled second-stage reagent to determine the location of the bound antibody and the corresponding epitope clone. Any labeled or tagged second-stage reagent useful for binding the bound antibody can in principle be incorporated into the procedure for the purposes of identifying the clones having epitopes bound by antibody. One preferred embodiment is soaking the washed disks having bound antibody in TTBS, 1% milk, 125I-protein A (0.5 to 1xcexc curie/ml) for 1.5 to 3 hours. Preferred ranges of conditions are incubating the disks for at least 1 hour, within a temperature range of between about 4xc2x0 C. to about 65xc2x0 C., in buffer near neutrality containing about 0.1% (v/v) to about 1.0% (v/v) neutral detergent, in about 0.1% to about 5% blocking agent and detectable quantities of labeled protein A. Another preferred second-stage reagent is labeled protein G, e.g., 125I-protein G. Other appropriate second-stage reagents include, but are not limited to, double antibody, such as 125I-labeled mouse anti-human IgG, or mouse anti-human IgG tagged with beta-galactosidase or peroxidase. Substantial purity of labeled second-stage reagent is desirable.
The disks having bound labeled antibody are now soaked or washed to remove unbound label. One preferred embodiment is soaking 20 minutes four times in TTBS. The location of the labeled, bound antibody on the disks is determined by conventional procedures appropriate for the labeled second-stage reagent. X-ray film is used for 125I. Chromogenic substrates are useful in a variety of enzyme-antibody detection kits.
Once the location of the bound antibody is determined, e.g., a pattern of dark spots on developed X-ray film, one identifies the appropriate colonies on the original mature plate, since the colonies are regrown as needed. Subsequent replating, growth, and sequencing gives a particular selected principal neutralizing epitope (SPNE).
For ease of evaluating SPNE as ligands, applicants have constructed recombinant shuttle vectors coding for RFPs of novel SPNE and selected peptides or fragments thereof, such as pIII (with or without a polyhistidine tail), Hep B core, Hep B surface-antigen or protein A. The methods for construction of fusion peptides are well known in the art. Coding sequences are prepared by ligation of other sequences, cloning, PCR, mutagenesis, organic synthesis, or combination thereof, in accordance with the principles and practice of constructing DNA sequences.
Once selection of desired phage epitopes in a phage epitope library has been made, it is necessary to determine the DNA sequences coding for a selected SPNE. The present invention utilizes PCR to amplify the SPNE and sequencing of the resulting fragment.
In particular, after one or more rounds of selection, E. coli colonies are grown overnight at about 37xc2x0 C. in a suitable medium containing appropriate antibiotics. The supernatant is used as template in PCR reactions. The template is amplified using 100-fold excess of one primer over the other. Template and oligonucleotide primers (Primer 1008: 5xe2x80x2-TCG AAA GCA AGC TGA TAA ACC G-3xe2x80x2 SEQ ID NO:1, located 106 nucleotides upstream of random insert and Primer 1009: 5xe2x80x2-ACA GAC AGC CCT CAT AGT TAG CG-3xe2x80x2 SEQ ID NO:2,
located 87 nucleotides downstream from random insert) are reacted in a volume of 100 xcexcl containing KCl; Tris-HCl, about pH 8.3; MgCl2; gelatin; each DNTP and a thermalstable DNA polymerase, e.g., Taq and others appreciated by those of ordinary skill in the art. Mineral oil is placed over the reaction and amplification in a thermal cycler is carried out for an initial period at about 94xc2x0 C. incubation, then about 30 cycles of about 30 seconds at about 94xc2x0 C., about 1 minute at about 55xc2x0 C. and about 2 minutes at about 72xc2x0 C. followed by about 5 minute incubation at about 72xc2x0 C. In a preferred embodiment, the mineral oil is removed, water added to the reactions, and the sample is centrifuged in a microconcetrator. The retentate volume is brought up to about 2 ml with water and centrifuged. The retentate is then collected by centrifugation. Retentate concentrations are determined by electrophoresis on an agarose gel containing Ethidium bromide and visualization under ultraviolet light. The retentate is dried along with enough limiting primer from PCR reaction (or internal primer 1059 5xe2x80x2-GTA AAT GAA TTT TCT GTA TGA GG-3xe2x80x2 SEQ. ID NO:3, located, 27 nucleotides downstream from insert) to give about a 5:1 primer:template molar ratio. The DNA/primer mixture is resuspended in water and Trisxe2x80xa2Buffer. The primer and template are annealed and chain termination sequencing reactions are set up and run. A sequencing gel is run on the PCR product. The gel is dried and exposed to X-ray film overnight and the sequence is then determined. Alternatively, other methods of sequencing the PCR product may be used, e.g., chemical cleavage, automated fluorescence sequencing described by Tracy, T E and L S Mulcahy, Biotechniques, 11, 68 (1991) or modifications thereof.
For the particular RFPs of this invention, DNA sequences coding for a selected SPNE are ligated in frame to DNA sequences coding for pIII, Hep B core or protein A. The resulting DNA fragment is expressed in any one of a wide variety of readily available systems, e.g., E. coli BL21(DE3), as also discussed later.
The HIV/pIII fusion was expressed in E. coli using the T7 polymerase system from Rosenberg, A. H. et al., Gene 56, 125 (1987). The plasmid pET-3a (commercially available from Novagen, Madison, Wis.) was digested with Xba I and BamHI and the 5 kb vector fragment isolated. The isolated vector fragment was ligated with the Xba I, BgI II-digested HIV-pIII fusion prepared by polymerase chain reaction (PCR) of the candidate HIV fusion phage clones.
Two synthetic DNA oligomers were used to amplify a portion of the phage pIII gene (including the HIV sequence) and append sequences which permit efficient expression and purification of the pIII product. The first synthetic DNA oligomers, 5xe2x80x2-CCC TCT AGA AAT AAT TTT GTT TAA CTT TAA GAA GOA GAT ATA CAT ATG GCC GAC GGG GCT-3xe2x80x2 (SEQ ID NO: 4), has homology with the fuse phage III gene with the sequences encoding the mature amino terminus of Ala-Asp-Gly-Ala (SEQ ID NO:5). PCR amplification from this site incorporates the sequences encoding the mature pIII protein and rebuilds the pET-3a vector from the Xba I sit to the initiating methionine.
The second synthetic DNA oligomer, sequence 5xe2x80x2-CTC AGA TCT ATT AAT GGT GAT GGT GAT GAT GTA TTT TGT CAC AAT CAA TAG AAA ATT C-3xe2x80x2 (SEQ ID NO:6) encodes the reverse strand of the carboxyl-terminal portion of pIII ending with residues Cys-Asp-Lys-Ile (SEQ ID NO:7). PCR with this oligo rebuilds the fuse phage pIII gene up to the transmembrane domain and appends six histidine residues to the carboxyl-terminal isoleucine. The presence of the histidine residues facilitates purification of the pIII fusion protein by metal chelation chromatography (Hochuli, E. et al., J. Chromat. 411, 177 (1987) using nitrilotriacetic (NTA) resin (available from Qiagen, Chatsworth, Calif.).
Expression of the pIII fusion is obtained by transforming the expression plasmid into E. coli strain BL21 (DE3) (Rosenberg, A. H. et al., supra; U.S. Pat. No. 4,952,496; Steen, et al., EMBO J. %, 1099 (1986). This strain contains the T7 phage RNA polymerase gene under control of the lac operator/promoter. Addition of isopropylthio-galactoside (IPTG) at culture OD600=0.6-0.8 induces T17 RNA polymerase expression which transcribes pIII mRNA to high levels. This RNA is translated yielding pIII fusion protein which is harvested 3-4 hours post induction and chromatographed on NTA resin.
In the alternative, the fusion peptides can be made by synthetic organic means, although this method is limited by feasibility and by practicality to smaller fusion peptides. See also the section on organic synthesis of SPNE, above.
Selection of the ligand-target molecule system is critical for any affinity chromatography purification scheme. Knowledge of the kinetics of the ligand and target molecule interaction is important for rationally designing the steps, materials and solutions used in the purification of a given target molecule. Knowledge of the affinity constant (Ka) alone may not be adequate for optimal system selection. In particular, knowledge of the association constant (ka) and the dissociation constant (kd) is required to optimally design a given purification scheme. For immunoaffinity chromatography, this knowledge has been, until recently, elusive because of the difficulty in studying antibodyxe2x80x94antigen interactions. Several methods have been described to identify antigens that bind to a specific antibody. Such methods include, but are not limited to, agglutination reactions, precipitation reactions, immunoassays, immunofluorescence, fluorescence-activated cell sorting. These methods do not represent the primary interaction between antibodies and a given epitope, but, rather, depend on secondary interactions for detection. This, therefore, makes it impossible to determine exactly the ka and kd for a given antibodyxe2x80x94antigen interaction and ultimately renders an optimally designed purification scheme a misnomer.
Recently, however, quantitative analysis of molecular interactions in real time has been described (Karlsson, R., Michaelsson, A. and Mattsson, L. (1991). Kinetic analysis of monoclonal antibody-antigen interactions is possible with a new biosensor based analytical system. Journal of Immunological Methods, 145:229-240). This analysis relies upon surface plasmon resonance as a direct optical sensing technique, based on total internal reflectance, to study molecular interactions, e.g., antibodyxe2x80x94antigen, in real time.
In a real-time biospecific interaction, light is coupled resonantly into electric oscillations, or surface plasmons, at a metal surface, e.g., gold. Such oscillations give rise to a nonpropagating evanescent wave that extends from the metal surface into the sample solution, decaying exponentially as a function of distance. Macromolecular complexes formed at the metal-liquid interface, resulting in a change in refractive index of the liquid media at the interface, perturb the evanescent wave and alter the propagation characteristics of the plasoms. Changes in the propagation characteristics of the plasmons in turn alter the characteristics of the internally reflected light. Such changes are ultimately detected and quantitated by means of a diode array. This instrument uses a layer of gold modified with carboxylated dextran to provide a hydrophilic surface for immobilization of macromolecules, e.g., proteins, immunoglobulins and antibodies. It is possible to site-direct the immobilization chemistry for macromolecule immobilization. For example, activation of the dextran matrix using a mixture of N-ethyl-Nxe2x80x2-(3-diethylaminopropyl)carbodiimide (EDC) and N-Hydroxy-succinimide (NHS) produces NHS-esters on the matrix for reaction with primary amino-containing macromolecules. Other methods use hydrazine, to produce an active hydrazine matrix, and sulfo-m-maleimidobenzoyl-N-hydroxysulfosuccinrimde ester (sulfo-MBS), to produce an active sulhydryl matrix. After one of the reactants is covalently attached to the dextran matrix, the other is introduced in a flow passing over the surface. The resonance angle depends on the refractive index in the vicinity of the metal surface and is monitored continuously, thus allowing the association or dissociation of molecules from the sensor surface to be followed in real time. No labeling of the ligand or the target molecules is required.
In one embodiment of this invention, 447 antibody is covalently bound to the dextran matrix after less than five minutes of activation. Specifically, activation is accomplished using a continuous flow of HBS (10 mM HEPES, 0.15 M NaCl, 3.4 mM EDTA and 0.05% Surfactant P20), about pH 7.4, passing over the sensor surface. The carboxylated dextran matrix is then activated by injecting a solution containing EDC and NHS. The target molecule is then injected followed by a blocking agent such as ethanolamine to block remaining NHS-ester groups. The target molecule-dextran matrix is then conditioned with an acid, e.g., HCl. After this step, the sensor surface is ready for use. An immobilization level of about 10,000 to about 15,000 RU, with about 10,000 RU being preferred, corresponding to 10-15 ng/mm2 of 447 antibody is preferred.
The 447 antibody is used to capture antigenic RFPs. An analytical cycle consists of injecting a fusion peptide in supernatant from about 1.5 minutes to about 5 minutes, preferably for about 3 minutes, dissociation of the fusion protein in buffer flow and regeneration of the surface with 100 mM HCl for 3 minutes. Analytical cycles can be programmed and the entire analysis is completely automated.
The medium used in this invention may be, but is not limited to, HEPES, MEM, NCTC, IMDM and RPMI media. Any cell line that is capable of expressing antibody, receptors or any other target molecule of interest is included within this invention. Cells lines include but are not limited to, NS/O. The medium may be conditioned from about 1 to about 14 days, with 8 days being preferred. Cells and cell debris may be removed from the medium using centrifugation, filtration or other methods known in the art.
In a preferred embodiment of this invention, a NS/O cell construct is grown in IMDM medium supplemented with protein growth factors for 8 days. The antibody containing conditioned medium is filtered by passing the conditioned medium through a 0.1 xcexcm filtration device followed by passing the resulting medium through a 0.22 xcexcm sterile filtration device to remove the intact cells and cell debris.
Affinity chromatography provides a convenient method for preparing pure macromolecules. In particular, immunoaffinity chromatography takes advantage of the high affinity interaction between an antigenic peptide or polypeptide and its corresponding antibody. Purified antibody may be obtained using immunoaffinity chromatography wherein the antigenic peptide or polypeptide is coupled to an inert matrix and is used as a selective adsorbent for antibody isolation.
Inmunoaffinity chromatography comprises three principle steps; adsorption, washing and elution. Adsorption and elution are the most critical for success. Adsorption is the step wherein the target molecule is bound to the ligand. Adsorption is accomplished by contacting the sample containing the target molecule with the ligand bound to solid support matrix in a suitable medium within a column. Washing is the step wherein impurities present in the fluid volume of the column as well as those bound nonspecifically to the ligand are removed. Washing is accomplished by passing a volume of physiological buffer, such as phosphate buffered saline, about pH 7.2, through the column. The volume of buffer used in the washing step should not be so great as to result in target molecule loss but, on the other hand, not so limited so as not to remove impurities. Elution is the step wherein the target molecule is removed from the column by using a solvent that reduces the affinity of the target molecule to the ligand or the affinity of the ligand-target molecule complex to the solid support. Elution of an antibody coupled to the antigen may be accomplished by either a salt gradient, to change the pH; buffered step-gradient, to change the ionic strength; or other methods.
Proper selection of a solid support for the ligand is critical for specific adsorption. The ideal matrix should possess several characteristics including, macroporosity, mechanical stability, ease of activation, hydrophilicity, and inertness, i.e., low nonspecific adsorption. No matrix is ideal in all of these respects; the matrix is often determined empirically. Matrices commonly used by those skilled in the art include cross-linked dextran, agarose, polyacrylamide, cellulose, silica and poly(hydroxyethylmethacrylate). For immuno-adsorbents, beaded agarose is the preferred solid support by those skilled in the art due to its high adsorptive capacity for proteins, high porosity, hydrophilicity, chemical stability, lack of charge and relative inertness toward nonspecific adsorption.
Ligands may be physically adsorbed to matrices or covalently attached to polymeric matrices containing hydroxylic or amino groups by means of bifunctional reagents. Attachment usually requires two steps, activation of the matrix and coupling of the ligand to the activated matrix. Activated matrices are available commercially. The selection method for coupling the ligand to the matrix is dictated in part by the choice of matrix, and, in part, by the choice of ligand.
Most methods commonly used to immobilize peptide or polypeptide ligands are based on coupling of amino groups. The polypeptide ligand must be coupled in a manner that will not interfere with its ability to be recognized by the target molecule. Methods for activation and coupling, commonly used by those skilled in the art, include but are not limited to cyanogen bromide, bisoxirane, N-hydroxysuccinimide esters and divinlysulfone. For detailed procedures see, e.g., Axen et al., Nature (London) 214, 1302 (1967), Porath et al., Protides Biol. Fluids, Proc. Colloq. 18:401 (1970), and Porath et al., Nature (London) 238, 261 (1972). The preferred method for activating agarose matrices by those skilled in the art is with cyanogen bromide. This method is relatively simple and can be performed entirely in aqueous solutions.
For successful use of affinity chromatography, the polymer-bound ligand must be sufficiently distant from the polymer surface to minimize steric interference. This is accomplished by inserting an interconnecting link or spacer between the ligand and the matrix. There are two methods commonly employed for introducing spacers. First, the ligand may be prepared with a long hydrocarbon chain containing an amino group which will serve as the spacer. Second, the spacer may be bound directly to the matrix so that the ligand can be attached directly to these spacers. Types of spacers commonly used by those skilled in the art include but are not limited to cystamine, p-aminobenzoic acid, tyramine and p-hydroxy-mercuribenzoate.
The specific buffering conditions used for equilibrating the affinity column in preparation for sample application should reflect the specific properties of the interacting system being used. The nature of the buffer used, including its pH and ionic strength, should be optimal for the ligand-target molecule system. The target molecule sample applied to the column should be contained in the same buffer used to equilibrate the column. After sample application and adsorption, the column should be washed with the starting buffer to remove any unbound sample and any impurities. It is also common to then wash the column with buffers different from the starting buffer in order to remove nonspecifically adsorbed substances.
Elution of the target molecule may be accomplished by a number of methods, including but not limited to these presented here. There are no covalent bonds involved in the interaction between antibody and antigen. Thus, the conditions of the buffer may be changed such that the affinity of the antigen-antibody complex falls sufficiently to destroy effective binding to each other or to the solid support. This is achieved by altering the pH, or the ionic strength of the buffer or both, or by chaotropic ions, e.g., cyanates. Increased separation may be obtained by gradient elution. In the case of immunosorption, the binding of a polypeptide antigen and antibody complex may be so strong that more harsh elution conditions are necessary, such as the use of buffers which are very strongly acidic or basic. Such elution conditions may irreversibly denature the desired antibody or exacerbate antigen leakage. Other methods of elution include use of chaotropic agents such as KSCN; organic solvents, e.g., ethylene glycol, DMSO, or acetonitrile; denaturing agents, e.g., 8 M urea or 6 M guanine; electrophoretic elution; pressure induced elution and metal ion elution. Incomplete elution results in both loss of product and loss of column capacity. Ideally, the elution conditions should allow for complete elution of the product after one or two column volumes have passed through the column. The exact nature of the elution agent is dictated by the nature of the antigen-antibody interaction. Detailed discussions of affinity chromatography can be found in Affinity Chromatography: A Practical Approach, edited by Dean P. D. G., Johnson, W. S., Middle, F. A., Affinity Chromatography, Principles and Methods, as published by Pharmacia, (Pharmacia LKB Biotechnology, Uppsala, Sweden), and Immunoaffinity Purification: Basic Principles and Operational Considerations, Yarmush, M. L, et al., (1992) Biotech Adv., 10:412-446.
In one embodiment of the present invention, a column of a quadradridentate chelating adsorbant, Ni++-nitrilotriacetate-SEPHAROSE, is prepared and charged with the RFP containing a hexahistidine tail (SEQ ID NO:8). The column is equilibrated with a phosphate buffer, about pH 7.0, containing NaCl. Conditioned medium containing HIV-1 antibody is loaded on the column and is washed with about 5 column volumes of the same buffer. The RFP-bound antibody is washed with TWEEN 80 is sodium phosphate buffer, about pH 7.0, containing sodium chloride. The wash step is performed by passing about 2 column volumes of the buffer-TWEEN-sodium chloride solution followed by a stopped flow incubation and subsequent column washes. The antibody is eluted from the RFP with a gradient of NaCl. Alternatively, the antibody is eluted from the RFP with gradient of MgCl2.
Another embodiment of the present invention uses a modification of the method of Porath (Porath, J., Methods in Enzymology, 34, 13 (1974)) wherein a volume of CNBr-activated SEPHAROSE is incubated with the RFP containing a lysine cluster tail (Gly-Ala-Lys-Lys-Ala-Lys, SEQ ID NO:9) for about 8 to 16 hours at about 0xc2x0 C. to about 10xc2x0 C., preferably about 4xc2x0 C., in sodium borate buffer, about pH 8.5. A column of the coupled SEPHAROSE is prepared and equilibrated with phosphate buffer, about pH 7.0, containing NaCl. Conditioned medium containing HIV-1 antibody is loaded on the column and is washed with the same buffer. The RFP-bound antibody is washed with TWEEN 80 in sodium phosphate buffer, about pH 7.0, containing sodium chloride. The wash step is performed by passing at least one column volume of the buffer-TWEEN-sodium chloride solution followed by a stopped flow incubation and subsequent column washes. An additional wash step is performed in the same column with MgCl2 followed by at least one wash with sodium phosphate buffer, about pH 7.0. The antibody is step-eluted from the RFP using glycine, about pH 3.0, and neutralized to about pH 7.0 with Tris-HCl, about pH 8.0. Alternatively, the antibody may be step-eluted from the RFP using MgCl2 in acetic acid and neutralized to about pH 7.0 using Tris-HCl, about pH 8.0.
Yet another embodiment of the present invention uses a volume of SEPHAROSE activated with 2-Fluoro-1-methylpyridinium toluene-4-sulfonate (FMP) coupled with the RFP (sequence presented above) according to published methods (Ngo, T. T., Biotechnology, 4; 134 (1986). A column of the coupled SEPHAROSE is prepared and equilibrated with phosphate buffer, about pH 7.0, containing NaCl. Conditioned medium containing HIV-1 antibody is loaded on the column and is washed with at least one column volume of the same buffer. The RFP-bound antibody is washed with TWEEN 80 in sodium phosphate buffer, about pH 7.0, containing NaCl. The wash step is performed by passing volumes of the buffer-TWEEN-sodium chloride solution through the column followed by a stopped flow incubation and a subsequent column wash. A second wash step is performed in the same column with MgCl2 followed by a wash with sodium phosphate buffer, about pH 7.0. The product is step-eluted using MgCl2 in acetic acid and neutralized to about pH 7.0 using Tris-HCl, about pH 8.0.
While the foregoing specification teaches the principles of the present invention, with examples provided for the purposes of illustration, it will be understood that the practice of the invention encompasses all of the usual variations, adaptations, modifications, deletions or additions of procedures and protocols described herein, as come within the scope of the following claims and its equivalents.
Specifically, 447 antibody described here is useful as a broadly neutralizing monoclonal antibody against HIV. This xe2x80x9c447 antibodyxe2x80x9d binds to about 90% of all known HIV serotypes and neutralizes HIV. It was isolated from a human patient.
Other receptors such as cytokines, other antibodies, protein receptors, recombinant antigen mimics and other conformational epitope mimics can be prepared according to the processes of the present invention and are included within the scope thereof.